mediators and moderators

Steps in Testing Mediation

1.Is beta coefficient significant for path c ?2.Is beta coefficient significant for path a ?3.Is beta coefficient significant for path b ?4.Last step is a regression: Predictors are both recess and physical activity

Mediator•

A variable that helps explain the meaningful association between two variables that have an existing correlation, because the mediator variable arises from or is internal to the independent variable.

Schoolchildren who have a daily recess break behave better and are likely to learn more, according to a large study of more than 10,000 third-graders.

But many kids now have less free time and engage in fewer physical activities at school because, in response to the "No Child Left Behind Act" of 2001, many school districts cut time allotted to recess in order to focus on reading and math.

Third variables vs. Mediator variables•So what is the difference?

Conceptual

Does path c get weaker or non-significant when the mediator is in the regression?

If yes, evidence in favor of mediation prediction.

•Third variable•A variable that represents a plausible alternative explanation for an existing correlation between two variables

May imply the existing correlation is spurious (meaningless) because the two variables are not directly related, meaning manipulating the independent variable would have no effect on the dependent variable

Playing violent video games is associated with aggressive behavior very strongly among teenagers, but less strongly among young adults. What type of prediction is this?

Moderator

Suppose we find that physical activity is related to lower levels of depression symptoms, and the relationship is stronger for older adults. What kind of variable is age in this example?

Moderator

Playing violent video games is associated with aggressive behavior because people who play violent video games have more aggressive parents. These aggressive parents are more likely to allow their kids to play violent games and to accept aggressive behavior in their children. What type of predicted model is this?

Third variable

Suppose we find that frequency of flossing is a predictor of lower risk of death. What would be an example of a plausible:

Third variable -conscientiousness•Moderator -economic level of the country•Mediator -dental health

Mediator versus Third variable Differences

Third variables are external to the bivariate correlation (problematic - nuisance, not of interest to researchers•Mediators are internal to the causal variable (not problematic- great interest to researchers, mechanism.

Mediators are internal to the causal variable (not problematic-

great interest to researchers, mechanism.

Playing violent video games is associated with aggressive behavior because children model what they see in video games.What type of prediction is this?

mediator

Is this relationship stronger (or weaker for one sub-group compared to another?

moderator

Third variables are external to the bivariate correlation (problematic -

nuisance, not of interest to researchers•

Mediator variables•Potential explanations for whywe see this relationship? Possible mediators schoolchildren

• Physical activity• Being outside?• Inhibition skills

Mediator versus Third variable similarities

•Both involve multivariate research designs.•Both can be detected using multiple regression

•Understand different forms of relationship between variables that are related

•Third variable•Mediator - WHY?•Moderator - FOR WHOM?

•Mediator WHY?

•medium = in the middle = middleman